Speech is useful for retrieving context and following an idea before it disappears. It is less reliable for exact names, specialist language, figures, citations, ownership, and commitments. That is why every workflow on this site separates capture from review.
The public web editors provide live browser recognition and editable text. They do not upload recorded meetings, identify speakers, verify claims, create citations, or replace the approved systems your school or organization uses. Browser and operating-system providers control recognition availability and may control audio processing.
Choose Voice Typing for short capture. Choose Dictation for a named, longer document with local browser autosave, editing controls, session timing, and export. In both cases, review important text before anyone relies on it.